The full-field strain distribution and the evolution behavior during additive manufacturing through in-situ observation

Accepted Manuscript The full-field strain distribution and the evolution behavior during additive manufacturing through in-situ observation

Ruishan Xie, Yue Zhao, Gaoqiang Chen, Xin Lin, Shuai Zhang, Shurui Fan, Qingyu Shi PII: DOI: Reference:

S0264-1275(18)30309-5 doi:10.1016/j.matdes.2018.04.039 JMADE 3852

To appear in:

Materials & Design

Received date: Revised date: Accepted date:

16 November 2017 22 March 2018 13 April 2018

Please cite this article as: Ruishan Xie, Yue Zhao, Gaoqiang Chen, Xin Lin, Shuai Zhang, Shurui Fan, Qingyu Shi , The full-field strain distribution and the evolution behavior during additive manufacturing through in-situ observation. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Jmade(2018), doi:10.1016/j.matdes.2018.04.039

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT The Full-field Strain Distribution and the Evolution Behavior during Additive Manufacturing Through In-situ Observation Ruishan Xie a,b, Yue Zhao a,b, Gaoqiang Chen a,b, Xin Lin c, Shuai Zhang a,b, Shurui Fan a,b,Qingyu Shi a,b,* a

State Key Lab. of Tribology, Department of Mechanical Engineering, Tsinghua University, Beijing,

PT

100084, PR China Key Laboratory for Advanced Material Processing Technology, Ministry of Education, PR China

c

State Key Laboratory of Solidification Processing, Northwestern Polytechnical University, Xi’an

SC

RI

b

* Corresponding author: [email protected]

NU

710072, PR China

Abstract: Part distortion is a technical bottleneck in the field of metal additive manufacturing,

MA

which generally depends on the thermo-mechanical behavior the material experienced during the deposition process. However, the transient strain distribution and evolution behavior of

D

the additive manufactured part still remain unclear due to the lack of in-process observation

PT E

method. This study successfully obtained the continuous full-field strain of a Ti-6Al-4V thin-wall during the laser engineered net shaping (LENS) additive manufacturing process using Digital image correlation (DIC) method. The evolution characteristic of vertical strain

CE

and longitudinal strain of the material was primarily studied during the deposition process. The results shows that the longitudinal strain was found increases rapidly to tensile strain as

AC

the laser beam approaches, whereas the vertical strain decreases rapidly to a compressive strain and gradually transform to tensile strain. Both vertical strain and longitudinal strain were found accumulated and rose periodically when depositing multi-layers, which increases the distortion tendency of the deposited part. In situ measurement of the strain field in additive manufacturing process can be an effective verification for theoretic and computational studies, which also provides the possibility of controlling stress and distortion in real time. Keywords: Additive manufacturing; Strain; Distortion; In-situ observation; Digital image correlation; Titanium alloys. 1

ACCEPTED MANUSCRIPT

1 Introduction Additive manufacturing (AM, also known as 3D printing) has been undergoing extensive development in recent years [1-4].

Laser engineered net shaping (LENS) is one of

PT

the most promising technologies in the AM system at the current state of the art, which has been applied in the manufacturing of metal parts in the aerospace field [2, 5]. However, one

RI

of the critical technical bottlenecks faced by this technology is the residual stress and distortion, which would substantially influence the dimensional precision of parts and even

SC

lead to direct cracking and unavailability of parts in severe cases [6-10]. The final distortion of additive manufactured parts chiefly depends on the thermo-mechanical behavior the

NU

material experienced during deposition process. Thus, research on the transient mechanical behavior of parts during AM process has significance for the prediction and control of

MA

residual stress and distortion [11, 12]. Yang [11] experimentally measured the final distortion of the bottom surface of a thin substrate using a laser three-dimensional scanner to validate

D

the numerical simulation model. Afazov [12] evaluated the final distortion deviation of the

PT E

additive manufactured blade by comparing the deformed configurations with a reference model. However, the methods mentioned above can only be implemented after the AM process finished. To reveal the real-time distortion, Michaleris [13, 14] attempted to measure

CE

the out-of-plan distortion of the free end of a cantilevered substrate using a laser displacement sensor. Even though these previous efforts have provided information of distortion during the

AC

AM process. However, the most important limitation of these method can only reflect the distortion of the substrate with finite points rather than the additive manufactured parts. There has been no literatures on the real-time measurement of the distortion of the additive manufactured parts directly. Digital image correlation (DIC) method has been widely used in the experimental measurement of dynamic strain field due to its non-contact, high precision, and full-field characteristics [15]. Recently, some researchers applied DIC method to measure the in-situ strain field of the area adjacent to the weld joint [16-19]. Strycker [16] compared the strain measurement results obtained by the DIC method and the electrical strain gauges during the 2

ACCEPTED MANUSCRIPT welding process. Their findings indicated that the strain evolution measured by the two methods coincide if enough images were taken in the DIC method. Ocelík [17] and Guo [18] successfully observed the in situ strain field and dynamic evolution of the backside of the specimen during the laser treatment process and welding process using the commercial DIC system. To suppress the influence of the intense arc light, Chen [19] designed a unique optical

PT

illumination and filtering system. In addition, a high-temperature sustaining speckle patterns required by the DIC algorithm was special prepared. The combination of these two efforts,

RI

they successfully measured the in situ strain field 1 mm away from the fusion line. Unlike the welding process, AM has more complex technical characteristics, adding to difficulties in

SC

strain observation. For instance, due to the technical characteristic of “start from zero”, it has great difficulties to prepare random speckles on the surface using the traditional methods.

NU

Considering this, DIC method has not yet been adopted in the AM process, although it has been proved to be an effective approach to reveal the continuous full-field strain behavior in

MA

real time in welding.

To summarize, the transient thermo-mechanical behavior the material experienced

D

during the deposition process remain unclear due to the lack of in-process observation

PT E

method. This study attempted to measure the in situ strain field during the AM process on the basis of DIC method. Additionally, the strain evolution during the multi-layer deposition and cooling process was revealed.

CE

2 Experiments and Methods

AC

A MLS2000 semiconductor laser was used for additive manufacturing of a Ti-6Al-4V thin-wall. Argon shielding gas was applied to protect the surface from oxdisation during the manufacturing process. The major laser processing parameters are as follows: laser power 2000 W, beam diameter 5 mm, scanning speed 600 mm/min, powder feeding rate 500 g/h. An HS-UX50 160K high-speed camera was used for photography. The maximum resolution of the photos was 1280 × 1024. The time of exposure ranged from 4 to 20 µs. Images taken from DIC camera were processed using the software Vic-2DTM to calculate the strain field. Figure 1a shows the DIC strain real-time measurement device during AM process. During the deposition process, the temperature of the material in the vicinity of the molten 3

ACCEPTED MANUSCRIPT pool rose rapidly and its radiation luminescence effect strengthened, thus making the material close to the molten pool area brighter while the material far from the molten pool area darker. Therefore, both supplemental lighting and optical filtering were used. A fixed light source was used for supplemental lighting; an optical filter was installed in front of the lens of the high-speed camera. Since there exists difficulty in preparing speckles ahead of time on the

PT

part surface, the tough surfaces of the formed parts served as the natural “speckle” after proper processing.

RI

The steps for the measurement are as follows: (1) a thin-wall with length of 50 mm and width of 5 mm was fabricated on a 100 × 50 × 6 mm substrate using laser cladding deposition;

SC

(2) a tough surface as a natural “speckle” with its own unique texture formed on the sides of the parts when the shaping reached a certain height; (3) processing was suspended (~240 s)

NU

for on-line debugging of the measuring system to obtain excellent imaging effect; (4) deposition continued and the images of the part surface were acquired using the DIC

MA

high-speed camera; and (5) Vic-2D software was used for image processing and to obtain the strain field distribution. The measured strain in the following deposition process represented

D

the strain increment at the time of measurement relative to the reference moment. In our

PT E

experiment, the first image was set as reference image when the deposition continued. The strain was determined by comparing the current image with the reference one. In the current experiment, the strain already existed in the deposited part which we took as reference strain,

CE

was not considered in the current experiments. The subset size and step size were set to 19 pixels and 7 pixels, respectively, as the software recommended. The detection lower limit of

AC

strain for Vic-2d system is up to 0.001%. The strain measurement noise in the experiments was about ±0.04%, which was determined by conducting the in-situ observation with no externally applied load before the deposition process. In the DIC experiments, the materials below the deposition layer underwent the periodically and quickly laser heating and cooling cycles during the following multilayer deposition process, which was the typical characteristics of the thermal cycle in AM . Through the process was suspended for a very short time for on-line debugging of the DIC system, the materials underwent the same typical thermal cycles mentioned above during the following deposition. Thus, the common law of the strain evolution in this paper was same 4

ACCEPTED MANUSCRIPT with that in the continuous deposition, through the specific value of the strain might be a little different. This study explores the strain distribution and evolution behavior of the parts when depositing two layers. The deposition direction of the first layer was from right to left (-x direction). The deposition direction of the second layer was from the left to right (+x). The time of deposition was 5 s. The interlayer cooling time was 2.6 s. For conciseness, the

PT

coordinate system used in this paper is shown in Figure 1b. The position where deposition started is defined as the zero point. The strain along the scanning direction (x direction) is the

RI

longitudinal strain, and the strain along the height direction (z direction) is the vertical strain. This paper focus on the in plane strains of the thin wall, while the strain in y direction was not

SC

measured in this paper, since no pronounced distortion in y direction was observed in the present small-dimension thin wall in experiments. The specific positions of the points to be

NU

investigated in the following sections are also presented in Figure 1b. P1 (x = -5), P2 (x = -25), and P3(x = -45) are three points 5 mm below the deposited layer parallel to the scanning

MA

directions. The time points at which the laser beam reached P1, P2, and P3 were 0.5 s, 2.5 s, and 4.5 s, respectively, when the first layer was deposited. P2, P4, and P5 are three points in

D

the middle (x = -25 mm): 5 mm, 10 mm, and 15 mm away from the molten pool, respectively.

PT E

The line segment Lx, which contains P1, P2 and P3, is located in 5 mm below the deposition layer. The line segment Lz, which contains P2, P4, and P5, is located in the center of the part.

CE AC

3 Results

Figure 1

3.1 Distribution characteristics of the strain field Figure 2a and 2b are the distribution fields of the longitudinal strain and vertical strain of the part surface, respectively, at t = 2.2 s when the first layer was deposited. As can be seen from Figure 2a, the maximum longitudinal strain (the red area) at this time point occurred at the right edge behind the molten pool and the strain in this area was positive, whereas the other areas were green and sky blue, i.e. the longitudinal strain was small. Figure 2b shows the cloud chart for vertical strain. A significant negative strain area (bluish violet) occurred in front of the molten pool; the strain of the area behind the molten pool gradually increased to 5

ACCEPTED MANUSCRIPT positive (yellow and red) whereas the strain in the other areas was small (green and sky blue). To investigate the distribution characteristics of the strains before and behind the molten pool, the strains along the line 5 mm below the deposition ( Lx in Fig. 1b) were extracted in Figure 2c. The distribution curve of longitudinal strain (red line in Figure 2c) indicates that the longitudinal strain was zero in the front and vicinity of the molten pool, and it gradually

PT

increased to 0.38% at the right edge with increase of the distance to the molten pool. The distribution curve of vertical strain (black line in Figure 2c), indicates that the vertical strain

RI

of the area at 14 mm before the molten pool (x = -36 mm) gradually decreased to a negative value from zero; and it decreased to the minimum (-0.47%) at 5 mm before the molten pool

SC

(x = -27 mm). Subsequently, the vertical strain increased gradually and transformed to a positive value in the vicinity of the molten pool; and it continuously increased behind the

NU

molten pool and reached 0.62% at the right edge. It is thus clear that the longitudinal strain

discussed in the following paragraphs.

MA

and vertical strain of the material below the deposited layer had differences, which are

Figure 2

PT E

D

3.2 Evolution history of strain field

Figure 3 is the curve of longitudinal strain evolution at different observation points parallel to the deposition direction. The blue bar in the figure represents the deposition stage

CE

and the red bar represents the interlayer cooling stage. The time points when the laser beam moves above various observation points are also indicated with arrows in the lower part of

AC

the figure. The deposition direction of the first layer is from right to left (-x direction). When the laser beam moved to the vicinity of point P1 at t = 0.5 s, the longitudinal strain increased rapidly. Then, it reached 0.42% after 1.6 s and decreased slowly afterwards. The longitudinal strain increased rapidly again when the laser beam for second layer deposition returned to the vicinity of P1 (~11.6 s) in the reverse direction. Likewise, the laser beam moved to the vicinity of points P2 and P3 at t = 2.5 s and t = 4.5 s in sequence. The longitudinal strain both increased to approximately 0.16%. When the second layer was deposited (+x direction), the laser beam reached P3, P2, and P1 in sequence. The longitudinal strains of the points increased to 0.53%, 0.32%, and 0.62%, respectively. In addition, after the laser beam passed, 6

ACCEPTED MANUSCRIPT the maximum strains of P1 and P3 at edges on both sides were 0.62% and 0.53%, which were significantly higher than the maximum strain of P2 in the middle (0.32%). This is because material constraint was smaller at the edges of the sides, while the material constraint in the middle was larger in the direction of parallel scanning. It can be inferred that the outward distortion trend of the material at the edge below the molten pool was large during laser

PT

deposition of the thin-wall. Figure 3

RI

Figure 4 presents the curve of longitudinal strain evolution at points P2, P4, and P5. The strains at different distances below the deposited layer exhibited a similar trend: when the

SC

laser beam move to the observation points, the longitudinal strain increased gradually; when the laser beam passed, with the increase in time the strain increased to a maximum and then

NU

decreased. As there was a certain distance between the observation point and the molten pool, the observation point far from the molten pool would experience delayed heat transfer. The

MA

longitudinal strains at P4 and P5 showed no significant signs of decrease during the cooling process.

D

Figure 4

PT E

Figure 5 presents the vertical strain evolution at P1, P2, and P3 in the parallel deposition direction. The vertical strain of P1 decreased rapidly to a negative value (-0.42%) when the first layer was deposited; after the laser beam left, its vertical strain increased gradually to a

CE

positive value with a maximum value of 0.68%. the vertical strain decreased to 0.58% during the stage of interlayer cooling. Likewise, the vertical strains at point P2 (in the middle) and

AC

point P3 (on the left) both decreased to negative strain values in sequence (-0.65%) before the laser beam moved to the observation points, and increased to ~0.4% after the laser beam move away. When the second layer was deposited (in reverse direction), the vertical strains at P3, P2, and P1 exhibited a variation trend similar to that during the process of first layer deposition. Specifically, the vertical strains first decreased and then increased to a higher level. Figure 6 presents the evolution history of the vertical strains at P2, P4, and P5. The variation trend of the vertical strain at P4 was similar to that at P2; while its variation amplitude was smaller than that at P2. The vertical strain at P5 (a point far from the molten 7

ACCEPTED MANUSCRIPT pool) 15 mm below the deposited layer varied a little (less than 0.2%). Figure 5 Figure 6 Based on a comparison between Figure 4 and Figure 6, the vertical strain was significantly higher than the longitudinal strain at the same position, which was largely

PT

associated with the constraint condition of the material below the molten pool in the x and z directions. The part may suffer from relative free distortion in z direction, since there was no

RI

addition constraint above the part. While the material distortion was constrained by material on both sides in x direction. In addition, both the vertical strain and the longitudinal strain of

SC

the area close to the existing deposited layer were greatly influenced by the deposition process. However, the strain in the area far from the existing deposited layer was less

NU

influenced by subsequent deposition. For instance, both the longitudinal strain (Figure 4) and the vertical strain (Figure 6) at positions 15 mm below the deposited layer were small and

MA

their variations were ignorable.

PT E

D

4 Discussion

In this paper, the laser beam interaction with the thin wall has high power density and moving velocity, and the spot size of the laser beam was small compared with the size of the wall.

CE

This case can be treated as a rapidly moving high-power line source interaction with the infinite plate in welding [20]for a simple analytical description. The temperature field of the

AC

deposited part could be approximately calculated by: T(r, t) =

2𝜂𝑞 𝑣ℎ(4𝜋𝜆𝑐𝜌𝑡)

𝑟2 2α exp (− + 𝑡) + 𝑇0 1 4𝑎𝑡 𝑐𝜌𝛿 2

(1)

Where 𝜂 is the the absorption efficiency, q is the laser power, 𝑣 is the scanning speed of the laser beam, ℎ is the thickness of the thin-wall, 𝑟 is the distance between the heat source and the observation point, 𝑇0 is the initial temperature, 𝜆 is the thermal conductivity, 𝑐 is the specific heat, α is convection coefficient. 𝑎 is the thermal diffusivity given by 𝑎 = 𝜆/𝑐𝜌, where 𝜌 is the density. In this paper, 𝑣=10 mm/s, ℎ =5 mm, q=2000 w, 𝑇0 =20 ℃. Other parameters are from reference [11]: 𝜂 =0.45, 𝜆 = 13.9e-3 w/mm/℃，𝑐 =0.682 J/g/℃, 𝜌 = 8

ACCEPTED MANUSCRIPT 4.43e-3 g/mm3, α = 55e-6 J/mm2/℃. Figure 7 shows the calculated temperature history of the deposited part during deposition process. It indicates that the heating rate and the peak temperature of the materials varies with the distance between the observation point and the laser beam, which decreased as the distance increased. The temperature of the materials near the melt pool increased rapidly to the peak value as the laser beam reached the observation

PT

locations, and it followed by a rapid cooling as the laser beam moved away. In addition, it can be found that the longer the distance from the laser beam, the longer it takes to reach the peak

RI

temperature.

SC

Figure 7.

It should be noted that the strain measured by DIC in the experiment was a macroscopic

NU

measurable strain, which was the total strain increment at the time of measurement relative to that in the beginning. The total strain increment composed of thermal strain increment, elastic

MA

strain increment, plastic strain increment, and the phase transformation strain increment. This paper discuss the evolution of the total strain during additive manufacturing, while the evolution of each strain component was hard to describe using DIC method. The strain

PT E

(1) Deposition stage

D

evolution behavior at different stages during additive manufacturing are discussed below.

Based on the discussion so far, it can be seen that the distribution and evolution of the

CE

longitudinal strain and the vertical strain exhibited different characteristics. During deposition stage, a large temperature gradient occurred in the area adjacent to the molten pool in the

AC

height direction due to the transient laser heating[21]. The material far below the deposited layer (all observation points in Figure 3-6) suffered from compression arising from the rapidly thermal expansion of the material adjacent to the molten pool. Thus, the vertical strain (Figure 5 and Figure 6) decreased rapidly as the laser beam approached the observation point. In terms of longitudinal strain, the temperature gradient parallel to the deposition direction at this point was relatively small; the compression effect in longitudinal direction from the adjacent material was limited. Thus, no significant decrease in longitudinal strain was observed when the laser beam passed (Figure 3 and Figure 4). When the laser beam passed the observation point, heat was continuously transferred to the observation point from the

9

ACCEPTED MANUSCRIPT area adjacent to the molten pool. The material temperature at this point increased gradually and the longitudinal strain and vertical strain also increased gradually. Similar strain evolution history was also observed via DIC method in welding, when the heat source approaches the observation point adjacent to the weld zone [17, 18, 22]. Qiao et al. [22] observed the transverse strain of the material on either side of the weld zone was

PT

initially a negative strain during laser welding, which finally became a positive transverse strain with the gradual cooling of the material. Ocelík [17] also observed the occurrence of

RI

compressive strain at the bottom of the plate when performing laser surface processing. The reason was attribute to the inhomogeneous heating of the material: the temperature at the

SC

upper laser treated surface is much higher than the temperature at the lower observed surface. Therefore the upper surface was expanding more strongly than the lower surface, leading to a

NU

negative strain at the bottom. In this paper, the obtained strain evolution history when depositing the first layer during AM were consistent with that in welding [17, 18, 22], which

MA

has the similar processing characteristics with steep thermal cycles. It indicates that the in-situ strain measurement method in this study had good reliability.

D

(2) Cooling stage

PT E

The cooling process is divided into two stages: the cooling process after the laser beam passes and the interlayer cooling process after each layer is deposited. During the cooling process, the temperature declines and the thermal strain decreased; while both the measured

CE

strains at P3 and P5 increased instead. This is associated with the cooling time interval and the distance to the currently deposition layer. The deposition direction is from right to left for

AC

the first layer, while it is in reverse direction for the second layer. The deposition direction of the second layer is from left to right. As for the first interlayer cooling, the time intervals for the laser beam to pass the P1 (right), P2 (middle), and P3 (left) observation points are ∆𝑡1 > ∆𝑡2 > ∆𝑡3 , in sequence. Thus, the long cooling time at the right P1 point makes both the longitudinal strain and the vertical strain decrease significantly during cooling stage. No significant decrease occurred in the strains of P2, due to the short cooling time and the hysteresis of material heat transfer. The strains of P3 increased instead during the interlayer cooling stage, which has the shortest cooling time. The strain evolution of the material also varies with the distance to the currently 10

ACCEPTED MANUSCRIPT deposition layer during cooling stage. Figure 7 indicates that the longer the distance from the deposition layer, the longer it takes to reach the peak temperature. It is because it takes time for thermal diffusion from the laser beam to the observation locations. During the interlayer cooling stage, the strains of the materials relatively close to the deposition layer (P2, z=-5 mm) shows a little decrease. However, as for the materials far from the deposition layer (P5,

PT

z=-15 mm), its strains increased instead because its temperature might keep increasing during the cooling stage.

RI

(3) Deposition of the second layer

When the second layer was deposited in reverse direction, the evolution history of

SC

longitudinal strain and vertical strain were similar to those of the deposition of the first layer. During multi-layer deposition, both vertical strain and longitudinal strain were found

NU

accumulated and rose periodically. It is largely due to temperature increase

and essentially

due to short interlayer cooling time[6]. The temperature of the formed parts did not decrease

MA

to room temperature after a short time of interlayer cooling (2.6 s in this study). Thus, both vertical strain and longitudinal strain repeated the similar evolution law of the first layer

D

deposition while maintaining the existing positive strains, instead of decreasing to the initial

PT E

(zero strain) state before the deposition of the second layer. The accumulation of transient strain in the reciprocating scanning mode would increase the distortion tendency in additive manufactured part.

CE

This paper made a very basic attempt to show the possibility of measuring the strain field directly on the additive manufactured part during the additive manufacturing process by

AC

using DIC method. The additive manufacturing process in this paper was not completely the same with the actual deposition process, which has more complex process with thousands/hundreds continuous deposition passes. Nevertheless, this paper got new insight to the strain evolution behavior of the deposited part during the following two-layer-deposition process, which involves the typical repeated and steep thermal cycles induced by the periodically and quickly laser heating and cooling in additive manufacturing. The common law of the strain evolution was same with that in actual deposition process, though the specific value of the strain might be a little different. Future work is undertaken to study the transient strain evolution during the whole additive manufacture process through both DIC 11

ACCEPTED MANUSCRIPT measurement and numerical simulation approaches.

5 Conclusions (1) The continuous full-field strain of a Ti-6Al-4V thin-wall during additive manufacturing was successfully measured base on DIC method with the rough surface as the

PT

natural speckle in combination with supplementary lighting, optical filtering, and high-speed photographing technologies.

RI

(2) During the additive manufacturing of a thin-wall, the vertical strain of the material

SC

below the deposited layer was found significantly greater than the longitudinal strain. The area closer to the molten pool had greater strain. Both the longitudinal strain and the vertical

NU

strain at 15 mm from the existing deposited layer can be ignored. (3) In the laser deposition stage, with the approach of the laser beam, the longitudinal

MA

strain of the parts in the area adjacent to the molten pool increased rapidly, while the vertical strain rapidly decreased to a negative value and gradually increased to a positive value after

D

the laser beam passed. During the interlayer cooling stage, both the longitudinal strain and

PT E

vertical strain of the material adjacent to the molten pool decreased with increasing cooling time. No significant decrease occurred in the strains of the areas with small intervals of cooling time or at a long distance from the molten pool.

CE

(4) During laser multi-layer deposition, both vertical strain and longitudinal strain were found accumulated rose periodically. The evolution history of the strains arising from the

AC

deposition of various layers were similar. The maximum longitudinal strain of the material at the edges below the deposited layer was significantly greater than that at the middle.

Acknowledgment This work was supported by the National Key Technologies R&D Program, China (Grant no. 2016YFB1100100).

12

ACCEPTED MANUSCRIPT

References [1] T. DebRoy, H.L. Wei, J.S. Zuback, T. Mukherjee, J.W. Elmer, J.O. Milewski, A.M. Beese, A. Wilson-Heid, A. De, W. Zhang, Additive manufacturing of metallic components – Process, structure and properties, Progress in Materials Science 92 (2018) 112-224.

PT

[2] W.E. Frazier, Metal Additive Manufacturing: A Review, Journal of Materials Engineering and Performance 23(6) (2014) 1917-1928.

RI

[3] W.J. Sames, F.A. List, S. Pannala, R.R. Dehoff, S.S. Babu, The metallurgy and processing science of metal

SC

additive manufacturing, International Materials Reviews 61(5) (2016) 315-360.

[4] S.K. Everton, M. Hirsch, P. Stravroulakis, R.K. Leach, A.T. Clare, Review of in-situ process monitoring and

NU

in-situ metrology for metal additive manufacturing, Materials & Design 95 (2016) 431-445. [5] F. Arcella, F. Froes, Producing titanium aerospace components from powder using laser forming, The

MA

Journal of the Minerals Metals & Materials Society 52(5) (2000) 28-30. [6] T. Mukherjee, V. Manvatkar, A. De, T. DebRoy, Mitigation of thermal distortion during additive

D

manufacturing, Scripta Materialia 127 (2017) 79-83.

PT E

[7] M. Megahed, H.-W. Mindt, N. N’Dri, H. Duan, O. Desmaison, Metal additive-manufacturing process and residual stress modeling, Integrating Materials and Manufacturing Innovation 5(1) (2016). [8] Y. Chen, K. Zhang, J. Huang, S.R.E. Hosseini, Z. Li, Characterization of heat affected zone liquation

CE

cracking in laser additive manufacturing of Inconel 718, Materials & Design 90 (2016) 586-594. [9] Z. Wang, E. Denlinger, P. Michaleris, A.D. Stoica, D. Ma, A.M. Beese, Residual stress mapping in Inconel

AC

625 fabricated through additive manufacturing: Method for neutron diffraction measurements to validate thermomechanical model predictions, Materials & Design 113 (2017) 169-177. [10] B.A. Szost, S. Terzi, F. Martina, D. Boisselier, A. Prytuliak, T. Pirling, M. Hofmann, D.J. Jarvis, A comparative study of additive manufacturing techniques: Residual stress and microstructural analysis of CLAD and WAAM printed Ti–6Al–4V components, Materials & Design 89 (2016) 559-567. [11] Q. Yang, P. Zhang, L. Cheng, Z. Min, M. Chyu, A.C. To, Finite element modeling and validation of thermomechanical behavior of Ti-6Al-4V in directed energy deposition additive manufacturing, Additive Manufacturing 12 (2016) 169-177. [12] S. Afazov, W.A.D. Denmark, B.L. Toralles, A. Holloway, A. Yaghi, Distortion Prediction and

13

ACCEPTED MANUSCRIPT Compensation in Selective Laser Melting, Additive Manufacturing

(2017).

[13] E.R. Denlinger, J.C. Heigel, P. Michaleris, T.A. Palmer, Effect of inter-layer dwell time on distortion and residual stress in additive manufacturing of titanium and nickel alloys, Journal of Materials Processing Technology 215 (2015) 123-131. [14] J.C. Heigel, P. Michaleris, T.A. Palmer, In situ monitoring and characterization of distortion during laser

PT

cladding of Inconel® 625, Journal of Materials Processing Technology 220 (2015) 135-145. [15] W. Yan, F. Lin, Self-correction method of out-of-plane motions in two-dimensional digital image

RI

correlation, Proceedings of the Institution of Mechanical Engineers, Part B: Journal of Engineering Manufacture (2016) 095440541667368.

with the Digital Image Correlation Technique, Welding Journal

SC

[16] M.D. Strycker, P. Lava, W.V. Paepegem, L. Schueremans, D. Debruyne, Measuring Welding Deformations (2011).

Coatings Technology 203(20-21) (2009) 3189-3196.

NU

[17] V. Ocelík, J. Bosgra, J.T.M. de Hosson, In-situ strain observation in high power laser cladding, Surface and

MA

[18] N. Guo, X. Yin, J. Liang, N. Ma, Weld Bead Distortion of Thin-Plate Using Weak Digital Image Correlation Method, Journal of Materials Engineering and Performance 25(11) (2016) 4952-4958.

D

[19] J. Chen, X. Yu, R.G. Miller, Z. Feng, In situ strain and temperature measurement and modelling during arc

PT E

welding, Science and Technology of Welding and Joining 20(3) (2014) 181-188. [20] D.I.H.D. Radaj, Heat Effects of Welding, Springer Berlin Heidelberg, New York, 1992. [21] L. Wang, S. Felicelli, Y. Gooroochurn, P.T. Wang, M.F. Horstemeyer, Optimization of the LENS® process

CE

for steady molten pool size, Materials Science and Engineering: A 474(1-2) (2008) 148-156. [22] Q. Dong-xiao, In-situ strain measurement near weld pool using digital image correlation, Electric Welding

AC

Machine 43(9) (2013) 1-4.

14

ACCEPTED MANUSCRIPT Highlight: 

In-situ strain observation during additive manufacturing process was performed via DIC method.



The continuous full-field strain distribution of a Ti6Al4V thin-walll during the process was obtained.

CE

PT E

D

MA

NU

SC

RI

PT

The strain evolution behavior during the multi-layer deposition was clarified.

AC



15

Graphics Abstract

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7